Read on Twitter
Latest out! Self-initiated behavioural change and disease resurgence on activity-driven networks. Work done with @gozzi_nicolo Martina Scudeler @danielapaolotti @a_baronca https://arxiv.org/abs/2011.03754 . A thread below.
What and Why? We consider a population that has experienced a first wave of infections that was stopped early through strong interventions and did not develop a significant immunity to prevent a second wave (i.e. resurgence).
As restrictions are lifted individuals implement behavioural changes to minimize the risk of infection. Very familiar unfortunately....
How? We model such a scenario considering a SIR epidemic model unfolding on activity-driven networks ( https://www.nature.com/articles/srep00469)
We describe self-initiated behavioural changes as variations either in the social propensity of individuals (as previously done by @alerizzo_ et al https://journals.aps.org/pre/abstract/10.1103/PhysRevE.90.042801), activity, or in the probability of interacting with people outside tight social groups, communities.
We explore the role of compliance considering a fraction of individuals implementing such behavioural measures selected either at random or as a function of their activity. In fact, maybe due to the nature of their work, the most active people cannot easily adapt
We first characterize the spreading unfolding on the simplest version of activity-driven networks where interactions are memoryless and random and derive the analytical expression for the basic reproductive number as a function of the different compliance.
Our results highlight the key importance of accounting for heterogeneity in activation patterns: the efforts of a large fraction of the network might be vane if the most socially active individuals do not adapt their behaviours.
We consider also a more realistic link creation mechanism introducing communities ( https://www.nature.com/articles/s41598-018-20908-x). Here self-initiated changes in behaviours either reduce nodes' activity or the interactions across communities (i.e. social bubbles).
Our results indicate that, while both the reduction of activity and inter-community links might reduce the impact of the disease, imperfect social bubbles are less efficient in stopping the virus
Limitations are many! 1) Our contribution is only theoretical and it should not be indented as a precise representation of reality. This is crucial to stress during a pandemic...
2) We did not consider that connected populations, age-structured populations nor the complex nature of real self-initiated behavioural change (here two reviews https://royalsocietypublishing.org/doi/full/10.1098/rsif.2010.0142 and https://royalsocietypublishing.org/doi/full/10.1098/rsif.2016.0820).
3) We have limited ourselves to an exploration of the phase space rather than fitting the parameters using real data.
4) we have neglected high-order complex temporal dynamics of real time-varying networks (here a recent review https://www.sciencedirect.com/science/article/pii/S0370157320302489 by @fede7j @AlicePatania @_jgyou @lordgrilo and others)
